The quick development of information technology leads to fast increase in all kinds of data. Large, distributed databases have been more and more popular. Therefore, the KDD and Data Mining provide some new approaches to understand data, and the discovery of sequential patterns is an important branch in the KDD research.Some classic algorithms of sequence patterns mining have been proposed and the most focused on mining the complete set of frequent patterns, decreased the performance of space. The mining of closed sequence not only provides the same information, but also is more compact and effective.The work of this dissertation aims at the problems mentioned above. The main context is as follows:1. The most of the developed closed patterns mining algorithms mine patterns from short to long and from bottom to up, which is inherently costly in both time and space complexity when support threshold is lower or the sequential patterns is longer. A new algorithm InverCios mining closed sequential patterns with the minimal length restriction is presented, which mines patterns from long to short and top to down. The extensive performance study shows the InverCios significantly outperforms the previous algorithms in time and space complexity.2. The study is made on the solution to closed sequential pattern mining algorithms based on the pruned concept lattice, the construct algorithm of synchronal pruned concept lattice is proposed, which dynamically constructs concept lattice by inserting attributes, and implements the pruning by the Apriori properties in concept lattice construction. The closed sequential patterns mining based on pruned concept lattice reduces the size of search space, and the efficiency of mining algorithm is improved. |